Frequency stabilizing systems for electronic apparatus are well known and have been utilized for many different types of apparatus. Stabilizing requirements can, of course, vary depending upon the intended use of the apparatus, and the degree of achieved stability can likewise vary, at least in some instances, depending upon the particular apparatus stabilized.
With respect to devices such as atomic clocks or the like, stabilizing requirements are quite stringent and apparatus of this type present unique stabilizing problems which have heretofore precluded achievement of fully satisfactory frequency stability, at least for some purposes.
Atomic clocks of many configurations have heretofore been studied, used and/or commercially produced, and it has heretofore been suggested that such clocks include systems using so-called atomic or molecular beams. For a more complete discussion of frequency standard clocks, see "Frequency Standards and Clocks: A Tutorial Introduction" By H. Hellwig, NBS Technical Note 616 (Revised) March 1974, U.S. Department of Commerce, National Bureau of Standards.
The most frequently used and important example of atomic clocks is the cesium beam atomic frequency standard or clock. Devices of this type now form the basis for time service as well as for precision navigation and communication systems, and are also used as the primary standard for the unit of time. For a more detailed review of atomic frequency standards, see "Atomic Frequency Standards: A Survey" By H. Hellwig, Proceedings of the IEEE, Vol. 63, No. 2, February 1975, and "Clocks and Measurements of Time and Frequency" by Helmut Hellwig, WESCON Technical Papers, Vol. 20, Los Angeles, Calif., Sept. 1976. For a more detailed review of the cesium beam frequency standard in realizing the unit of time see "The Realization of the Second" By Helmut Hellwig, David W. Allen, Stephen Jarvis, Jr., and David J. Glaze, Proceedings of Fifth International Conference on Atomic Masses and Fundamental Constants (AMCO-5), Paris, France, June 2-6, 1975.
Atomic or molecular resonance devices, when used as frequency standards, have heretofore commonly utilized a cavity through which a beam of particles are directed. For such a cesium standard, a cesium oven emits a beam of cesium atoms which are coupled through a magnetic field, which acts as a state selector, to the cavity which acts as an interaction region for a microwave signal coupled thereto from an oscillator, normally referenced to a crystal oscillator. From the cavity, the beam is then directed through a second magnetic field, which acts as a sorting device, to a detector, such as an atomic detector. The oscillator is controlled by a feedback loop coupled from the detector.
The properties of the cavity to a high degree determine the performance of the device as a frequency standard or clock in terms of accuracy and long-term frequency stability. During the past approximately twenty years of development of cesium beam devices, many different cavity configurations and different modulation schemes have been tried in connection therewith. Most notably, these developed cavities have been of the so-called Rabi type, which have a single, uniform cavity structure, or of the so-called Ramsey type, which have two regions of interaction, that are spatially separated but still part of the same microwave cavity. For a discussion of the Ramsey type cavity see, N. F. Ramsey, Molecular Beams, Oxford University Press, London, England, 1956. Pulse excitation of atomic and molecular beam devices having Ramsey type interaction regions is discussed in "Evaluation and Operation of Atomic Beam Tube Frequency Standards Using Time Domain Velocity Selection Modulation" by H. Hellwig, S. Jarvis, Jr., D. Halford, and H. E. Bell, Metrologia, 9, 107-172 (1973).
In the cesium standard, the properties of the cavity affect the apparent frequency of the cesium resonance, i.e., the interaction of the cavity with the resonating cesium atoms may cause an apparent shift of the resonance frequency from the true resonance frequency of the atoms. This effect is called "the cavity phase shift" because it is caused by a non-uniform phase in the microwave cavity, either an end-to-end phase or a distribution of phases along and across the atomic beam trajectory in the cavity. This cavity phase shift currently limits the absolute accuracy of primary cesium standards to about 1 .times. 10.sup.-13 and the accuracy of commercial units to about 7 .times. 10.sup.-12. There is also evidence that a changing cavity phase shift (changing with time) causes long-term frequency changes and instabilities (over the period of months or years) in such devices, limiting their usefulness as clocks in time generation.
Since cavity phase distribution must be kept to very small tolerances in order to produce structures acceptable for performance, this is a significant contributor to the cost of the standards because of such factors as the required mechanical precision in the production of the cavity, the required testing of the cavity, and the rejection of atomic clocks due to cavity failure found in complete standards. It has been found that the attention in fabrication and testing that must be given to the cavity may account for up to 25% of the cost of the complete structure and 50% of the beam tubes. Further refinements of present machinery tolerances and further improvements of significance beyond that known at present, without a basic technological change, appear unlikely, and thus other solutions must be found if better stability is to be achieved.
Different electromagnetic field configurations (modes) have also been used in the cavities in an attempt to optomize the properties of the resulting standard. The modulation schemes for line-center lock which have been most generally employed have been of the frequency and phase-modulation type, but sine wave and square wave modulation have been also heretofore employed. While such servo systems have, at least in some instances, improved system stability, improvements in such systems have nevertheless still been needed.
Among prior art patents, the patents to Mainberger (U.S. Pat. No. 2,960,663), Zacharais et al (U.S. Pat. No. 2,972,115), and McCoubrey et al. (U.S. Pat. No. 3,060,384) relate to atomic resonance type devices with an electronic servo system, and the patents to Orenberg (U.S. Patent No. 3,042,878), Holloway et al. (U.S. Pat. No. 3,076,942) and Grant et al (U.S. Pat. No. 3,088,078) relate to frequency locking of a device that includes a Ramsey type cavity.
Thus, while attempts have been made to increase stability in beam type devices both by careful attention to cavity construction and by providing servo systems, achievement of even better stability for these devices is still needed to make such devices more fully suitable and acceptable for use as frequency standards.